Human-Machine Student Model Discovery and Improvement Using DataShop

Stamper, John C.; Koedinger, Kenneth R.

doi:10.1007/978-3-642-21869-9_46

Cited by 39 publications

(31 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [11], we presented a data-driven method for researchers to use data from educational technologies to identify and validate improvements in a cognitive model. For statistical modeling purposes, we used a simplification of a cognitive model made up of hypothesized components of knowledge or skills that students must acquire to be successful on target assessment tasks or activities.…”

Section: Using Educational Technology Data For Cognitive Task Analysismentioning

confidence: 99%

“…For statistical modeling purposes, we used a simplification of a cognitive model made up of hypothesized components of knowledge or skills that students must acquire to be successful on target assessment tasks or activities. These knowledge components (KCs) identify latent variables in a logistic regression model called the Additive Factors Model (AFM) [11], which is a generalization of item-response theory [12]. The method involves a wash-rinse-repeat iteration: 1) inspect learning curve visualizations and best-fitting parameters of AFM for a given set of knowledge components (a KC model), 2) hypothesize changes to the KC model based on identified problematic KCs, and 3) refit AFM with the new KC model and return to step 1.…”

Section: Using Educational Technology Data For Cognitive Task Analysismentioning

confidence: 99%

“…In our first attempt at redesign (briefly described in [11]), we identified four types of problems: unscaffolded, table scaffolded, area scaffolded, and problem statement scaffolded. Table scaffolded problems reflect the current setup in the tutor and include columns for intermediate areas (as in Fig.…”

mentioning

confidence: 99%

“…Problem statement scaffolded problems have the same table as the unscaffolded problems but provide an explicit hint in the problem statement directing the student to first find the component areas. During the implementation of this first redesign attempt [11], we experienced some issues with the parameter settings and knowledge tracing algorithm which resulted in students never mastering all skills. We also found that the problem statement scaffolded problems did not seem to help the students learn the KCs, so we removed this type of problem in the next design iteration.…”

mentioning

confidence: 99%

See 3 more Smart Citations

Using Data-Driven Discovery of Better Student Models to Improve Student Learning

Koedinger

Stamper

McLaughlin

et al. 2013

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. Deep analysis of domain content yields novel insights and can be used to produce better courses. Aspects of such analysis can be performed by applying AI and statistical algorithms to student data collected from educational technology and better cognitive models can be discovered and empirically validated in terms of more accurate predictions of student learning. However, can such improved models yield improved student learning? This paper reports positively on progress in closing this loop. We demonstrate that a tutor unit, redesigned based on data-driven cognitive model improvements, helped students reach mastery more efficiently. In particular, it produced better learning on the problem-decomposition planning skills that were the focus of the cognitive model improvements.

show abstract

Section: Using Educational Technology Data For Cognitive Task Analysismentioning

confidence: 99%

Section: Using Educational Technology Data For Cognitive Task Analysismentioning

confidence: 99%

mentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Using Data-Driven Discovery of Better Student Models to Improve Student Learning

Koedinger

Stamper

McLaughlin

et al. 2013

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

“…While estimation might be considered to be a KC in its own right, it is certainly an integrative KC (Stamper & Koedinger, 2011) composed of smaller KCs such as the abilities to round and perform basic arithmetic operations. The following sections describe in greater detail how estimation might be decomposed into individual KCs.…”

Section: The Kli Frameworkmentioning

confidence: 99%

Feedback in Software and a Desktop Manufacturing Context for Learning Estimation Strategies in Middle School

Malcolm

View full text Add to dashboard Cite

The ability and to make good estimates is essential, as is the ability to assess the reasonableness of estimates. These abilities are becoming increasingly important as digital technologies transform the ways in which people work. To estimate is to provide an approximation to a problem that is mathematical in nature, and the ability to estimate is closely related to several key mathematical skills and aptitudes. These combined aptitudes allow individuals to judge reasonableness, or the proximity of estimates to the actual exact answer. Types of estimation such as computational and measurement estimation hinge on this sense of the reasonableness of an answer, but the concept of reasonableness itself is not given sufficient attention in primary and secondary grades in the U.S.Digital technology provides the means to explore and refine student estimation strategies. Four studies were conducted with upper-elementary and middle school student participants. These featured the Estimation Calculator that provides immediate feedback on the percent error of estimates.The first three exploratory studies provided a backdrop for the fourth study, which was conducted at a public middle school. This study also featured emergent desktop manufacturing technology. Such technology allows digital designs to be printed in three dimensions. The geometric nature of 3D design was used in conjunction with estimation instruction for a unit on computational and volume estimation. The guiding question in the research was whether instruction and practice with estimation-feedback software and 3D design software would influence students' estimates and strategies.Estimation skills vary widely by class assignment. Students in a standard seventh grade mathematics class made estimates that were significantly less reasonable (partial ω 2 = .02) than those of two advanced classes. The complexity of their strategies, as measured by digits that were included or ignored during rounding, was notably higher than their peers, suggesting that they were unfamiliar with simplification in estimation (Lefevre, Greenham & Waheed, 1993). An interaction effect (partial ω 2 = .03) was present between the complexity of their estimates and those of their peers in advanced mathematics classes, suggesting an aptitude-treatment-interaction. Estimates using volume were hindered by computational estimation errors for all groups.

show abstract

Data mining and education

Koedinger

D’Mello

McLaughlin

et al. 2015

WIRES Cognitive Science

Self Cite

101

View full text Add to dashboard Cite

An emerging field of educational data mining (EDM) is building on and contributing to a wide variety of disciplines through analysis of data coming from various educational technologies. EDM researchers are addressing questions of cognition, metacognition, motivation, affect, language, social discourse, etc. using data from intelligent tutoring systems, massive open online courses, educational games and simulations, and discussion forums. The data include detailed action and timing logs of student interactions in user interfaces such as graded responses to questions or essays, steps in rich problem solving environments, games or simulations, discussion forum posts, or chat dialogs. They might also include external sensors such as eye tracking, facial expression, body movement, etc. We review how EDM has addressed the research questions that surround the psychology of learning with an emphasis on assessment, transfer of learning and model discovery, the role of affect, motivation and metacognition on learning, and analysis of language data and collaborative learning. For example, we discuss (1) how different statistical assessment methods were used in a data mining competition to improve prediction of student responses to intelligent tutor tasks, (2) how better cognitive models can be discovered from data and used to improve instruction, (3) how data-driven models of student affect can be used to focus discussion in a dialog-based tutoring system, and (4) how machine learning techniques applied to discussion data can be used to produce automated agents that support student learning as they collaborate in a chat room or a discussion board.

show abstract

Human-Machine Student Model Discovery and Improvement Using DataShop

Cited by 39 publications

References 6 publications

Using Data-Driven Discovery of Better Student Models to Improve Student Learning

Using Data-Driven Discovery of Better Student Models to Improve Student Learning

Feedback in Software and a Desktop Manufacturing Context for Learning Estimation Strategies in Middle School

Data mining and education

Contact Info

Product

Resources

About