April Galyardt

Virtual Mentoring and Persistence in STEM for Students With Disabilities

The purpose of this study was to investigate the effectiveness of virtual mentoring for enhancing the persistence of secondary and postsecondary students with disabilities engaged in science, technology, engineering, and mathematics (STEM) learning. The student participants (N = 189) were all engaged in STEM coursework and enrolled in a virtual-mentoring program for a span of 4 years. Persistence was measured with an online survey designed to evaluate growth across self-determination, self-advocacy, STEM aspirations, and self-efficacy (math and science). After participating in virtual-mentoring practices, the most significant improvement was demonstrated in students’ perception of self-determination and self-advocacy. Growth differences were identified across type of disability and race/ethnicity populations. Implications pertaining to STEM aspirations and self-efficacy were addressed.

Viz-R: Using Recency to Improve Student and Domain Models

We describe a new method to troubleshoot and improve domain and student models from interactive learning environments. The method applies as long as the models can generate predictions of student behavior. The method is a visualization of model predictions, categorized using a metric of recent performance. We describe the method, its application in prior work to student models, and a proposed extension to domain models.

STEM Scalable Model for Enhancing Secondary and Postsecondary Student On-Line Services

The purpose of this paper is to examine the BreakThru e-mentoring model for scalability purposes. Two aspects of this STEM e-mentoring program were examined: (1) the use of virtual environments and social media settings; and (2) the development of e-mentoring relationships (i.e., quality and engagement). Three secondary and three postsecondary institutions participated in the project. Mentors (n = 33) were recruited from postsecondary faculty, secondary teachers, graduate students, and business leaders. Of the BreakThru participants (n = 188), 57 % of the students continued in the program for multiple years. Specific design issues are described as essential for developing and measuring the outcomes of a similar student on-line resource.

Evaluating Simplicial Mixtures of Markov Chains for Modeling Student Metacognitive Strategies

Modeling and discovery of the strategies that students use, both cognitive and metacognitive, is important for building accurate models of student knowledge and learning. We present a simulation study to examine whether simplicial mixtures of Markov chains (SM-MC) can be used to model student metacognitive strategies. We find that SM-MC models cannot be estimated on the moderately sized data sets common in education, and must be adapted to be useful for strategy modeling.