Enter your EdLab username.
Enter the password that accompanies your username.
Create new account    Request new password
Submitted by Sharon Hsiao on Thu, 2014-07-17 13:17

What is a flag for? Social media reporting tools and the vocabulary of complaint (2014) Kate Crawford & Tarleton Gillespie, New Media & Society, pp, 1-19, doi: 10.1177/1461444814543163

This paper is just published, and on my birthday, so I figure i read it.

Flagging is a common mechanism for reporting offensive content in social media, including Facebook, Twitter, Vine, Flickr, YouTube, Instagram, and Foursquare, as well as in the comments sections on most blogs and news sites. It's used to filter, report abuse, individual expressions of concerns, etc. it may seem like a ubiquitous mechanism of governance, but there are no direct or uncomplicated representation of community sentiment on it yet. The paper discusses issues on content regulation, user awareness on social platform, controversial examples can become opportunities for substantive public debates, etc. The paper aims to bring more clarity to the range of functions of flags play in social media. It's pretty interesting for researchers to know & to think, especially those who plan on using "engineered features" to quantify community values with practical and rhetorical values.

Submitted by Sharon Hsiao on Wed, 2014-07-16 13:02

One of the best papers from EDM14.

Full paper can be retrieve here

Parameterized exercises have recently emerged as an important tool for online assessment and learning. The ability to generate multiple versions of the same exercise with different parameters helps to support learning-by-doing and decreases cheating during assessment. At the same time, our experience with using parameterized exercises for Java programming reveals suboptimal use of this technology as demonstrated by repeated successful and failed attempts to solve the same problem. In this paper we present the results of our work on modeling and examining patterns of student behavior with parameterized exercises using Problem Solving Genome, a compact encapsulation of individual behavior patterns. We started with micro-patterns (genes) that describe small chunks of repetitive behavior and constructed individual genomes as frequency pro les that shows the dominance of each gene in individual behavior. The exploration of student genomes revealed that individual genome is very stable, distinguishing students from their peers and changing very little with the growth of knowledge over the course. Using the genome, we were able to analyze student behavior on the group level and identify genes associated with good and bad learning performance.

Submitted by Sharon Hsiao on Tue, 2014-07-08 06:58

Abstract–Forums play a key role in the process of knowl- edge creation, providing means for users to exchange ideas and to collaborate. However, educational forums, along several others online educational environments, often suffer from topic disruption. Since the contents are mainly produced by participants (in our case learners), one or a few individuals might change the course of the discussions. Thus, realigning the discussed topics of a forum thread is a task often conducted by a tutor or moderator. In order to support learners and tutors to harmonically align forum discussions that are pertinent to a given lecture or course, in this paper, we present a method that combines semantic technologies and a statistical method to find and expose relevant topics to be discussed in online discussion forums.

paper is available here

Submitted by Sharon Hsiao on Mon, 2014-07-07 09:14

Abstract– Computer programs are a specific type of know- ledge artefacts that result from a creative process under strong formal restrictions. From an educational perspective, it has been argued that programming supports intellectual develop- ment and knowledge building. In this paper, we give a short overview of a system created to automatically detect the crea- tivity of solutions to programming exercises that address gen- eral mathematical and algorithmic skills. A first step to make these artefacts susceptible to automatic analysis was the defini- tion a descriptive feature set that captures both structural and procedural aspects of each solution. Secondly, machine learn- ing techniques have been used to form higher-level metrics simulating expert judgments on a given set of solutions. It turned out that expert judgments of program creativity differ considerably and systematically. This also led to a classification of the experts.

Submitted by Sharon Hsiao on Mon, 2014-07-07 04:32

Recent research has provided evidence that conversational agents can effectively be used to trigger and scaffold peer discource in computer-supported collaborative learning (CSCL) settings. In this study, we use a prototype conversational agent system named MentorChat to explore the impact of two different intervention modes on inducing beneficial students' interactions, following the Academically Productive Talk (APT) perspective. We analyze the effects of (a) unsolicited agent interventions, which are automatically initiated and displayed by the agent, as compared to (b) solicited agent interventions, which are initiated automatically but only displayed upon students' request. The outcomes indicate that an unsolicited intervention mode can be more effective than a solicited one by means of increasing the level of explicit reasoning displayed in students' dialogues.

Submitted by Sharon Hsiao on Sun, 2014-07-06 03:09

Dr. Popović is the Director of Center for Game Science at the University of Washington and founder of Engaged Learning

Most of the current work on improving learning outcomes focuses on a small subset of variables of an immensely multi- dimensional space of the learning ecosystem. With ITS, learning games, and other digital content we consider only individual students, other research focuses only on teacher development, or only on curriculum improvement. In this talk I will describe our efforts on how to discover optimal parameters of this system that considers student factors (engagement and mastery), classroom factors (blended learning and group learning variations), curriculum factors (multidimensional variation of existing curricula), and teacher factors (in-class tools that mitigate weaknesses, and promote teacher development). I will describe our work on algorithms to discover optimal learning pathways in this high-dimensional space and conclude with recent outcome

Submitted by Sharon Hsiao on Sat, 2014-07-05 09:50


The growing prevalence of e-learning systems and on-line courses has made educational material widely accessible to students of vary- ing abilities, backgrounds and styles. There is thus a growing need to accomodate for individual differences in such e-learning systems. This paper presents a new algorithm for personliazing educa- tional content to students that combines collaborative filtering algo- rithms with social choice theory. The algorithm constructs a “dif- ficulty” ranking over questions for a target student by aggregating the ranking of similar students, as measured by different aspects of their performance on common past questions, such as grades, number of retries, and time spent solving questions. It infers a difficulty ranking directly over the questions for a target student, rather than ordering them according to predicted performance, which is prone to error.

Submitted by Sharon Hsiao on Sat, 2014-07-05 05:51

Another fascinating paper:
Can Choices Made in Digital Games Predict 6th-Grade Students' Math Test Scores?

In this paper, we mined students' sequential behaviors from an instructional game for color mixing called Lightlet. Students pkaying the game have two broad strategies. They can either test candidate color combinations in an experiment room without risking an incorrect answer. Or they can choose colors from a faux shopping Catalog containing several different mixing charts. While the results shown in the Experiment Room are always correct, only a few of the charts in the Catalog are correct. Thus, if students use the catalog students must apply critical thinking skills to determine what charts to trust. Our primary goal in this work was to identify the crucial choice pattern(s) in students' game play that would contribute to their learning or subsequent performance.

Submitted by Sharon Hsiao on Sat, 2014-07-05 05:26

Who's in Control?: Categorizing Nuanced Patterns of Behaviors within a Game-Based Intelligent Tutoring System

The authors use dynamical analyses to investigate the relation between students' patterns of interactions with various types of game-based features and their daily performance. High school students (n=40) interacted with a game-based intelligent tutoring system across eight sessions. Hurst exponents were calculated based on students' choice of interactions with four types of game- based features: generative practice, identification mini-games, personalizable features, and achievement screens. These exponents indicate the extent to which students' interaction patterns with game-based features are random or deterministic (i.e., controlled). Results revealed a positive relation between deterministic behavior patterns and daily performance measures. Further analyses indicated that students' propensity to interact in a controlled manner varied as a function of their commitment to learning. Overall, these results provide insight into the potential relations between students' pattern of choices, individual differences in learning commitment, and daily performance in a learning environment.

Submitted by Sharon Hsiao on Sat, 2014-07-05 04:58



Modeling and predicting individual behavior in such interactive environments is crucial to better understand the learning process and improve the tools in the future. A model-based approach is a standard way to learn student behavior in highly-structured systems such as intelligent tutors. Defining such a model relies on expert domain knowledge.

They propose a data-driven approach to learn individual behavior given a user's interaction history. This model does not heavily rely on expert domain knowledge. They use framework to predict player movements in two educational puzzle games, demonstrating that our behavior model performs significantly better than a baseline on both games. This indicates that our framework can generalize without requiring extensive expert knowledge specific to each domain.