Joseph Michaelis | University of Wisconsin-Madison (original) (raw)

Uploads

Papers by Joseph Michaelis

Making connections between graphical representations is integral to learning in science, technolo... more Making connections between graphical representations is integral to learning in science, technology, engineering, and mathematical (STEM) fields. However, students often fail to make these connections spontaneously. Intelligent tutoring systems (ITSs) are suitable educational technologies to support connection making. Yet, when designing an ITS for connection making, we need to investigate what concepts and learning processes play a role within the specific domain. We describe a multi-methods approach for grounding ITS design in the specific requirements of the target domain. Specifically, we applied this approach to an ITS for connection making in chemistry. We used a theoretical framework that describes potential target learning processes and conducted a series of four empirical studies to investigate what role graphical representations play in chemistry knowledge and to investigate which learning processes related to connection making play a role in students' learning about chemistry. These studies combined multiple methods, including knowledge testing, eye tracking, interviews, and log data analysis. We illustrate how our findings inform the design of an ITS for chemistry: Chem Tutor. Results from two pilot studies done in the lab and in the field with altogether 99 undergraduates suggest that Chem Tutor leads to significant and large learning gains on chemistry knowledge.

The development of literacy and reading proficiency is a building block of lifelong learning that... more The development of literacy and reading proficiency is a building block of lifelong learning that must be supported both in the classroom and at home. While the promise of interactive learning technologies has widely been demonstrated, little is known about how an interactive robot might play a role in this development. We used eight design features based on recommendations from interest-development and human-robot-interaction literatures to design an in-home learning companion robot for children aged 11–12. The robot was used as a technology probe to explore families' (N = 8) habits and views about reading, how a reading technology might be used, and how children perceived reading with the robot. Our results indicate reading with the learning companion to be a way to socially engage with reading, which may promote the development of reading interest and ability. We discuss design and research implications based on our findings.

A potential benefit of the proliferation of out-of-school STEM programs is increased interest in ... more A potential benefit of the proliferation of out-of-school STEM programs is increased interest in engineering, which targets the well-documented need for recruitment and retention of high quality entrants into the engineering workforce. Hidi and Renninger’s (2006) Four-Phase Model of Interest Development (FPMID) posits that development of a person’s interest requires proper support to trigger, develop, and maintain interest in a domain. According to FPMID, a student’s interest can be triggered temporarily by a highly engaging situation (situational interest), or from a personal predisposition to engage that is more enduring and intrinsically motivated (individual interest). Each developmental phase of interest requires appropriately matched forms of support and learning opportunities in order to maintain and foster a student’s developing interest. In this study we propose the use of Lave and Wegner’s (1991) Legitimate Peripheral Participation in Communities of Practice as a model for understanding the role that the socio-cultural environment plays in the development of an individual’s interest. Without differentiated support from the community, a person may regress in their interest in a domain, or lose interest entirely. Thus, to appropriately meet students’ needs, program developers need to understand the process of increasing and maintaining student interest and design programs to meet the interest development needs of students.
The goal of this paper is twofold. First, as an example of measuring program efficacy in developing interest in engineering, we report on the authors’ findings from the use of the Four-Phase Interest Development in Engineering Survey (FIDES 2.0) with high-school students in an out-of-school/after-school engineering program. Second, we provide a detailed description of the in situ process of interest maintenance and development as described by the FPMID through the lens of Legitimate Peripheral Participation for these after-school program participants.
Program efficacy was determined using FIDES 2.0, a reliable and validated instrument that assesses interest levels across a broad range of indicators. FIDES 2.0 was administered twice to 9th – 12th grade students (N=13) in an out-of-school robotics competition program in order to assess baseline interest and then accurately document changes in participants’ interest in engineering. Initial (Mean = 5.47, σ = 0.74, α = 0.81) and post-competition scale scores 20 weeks later (Mean = 5.79, σ = 0.67, α = 0.84) confirmed that FIDES 2.0 was a highly reliable instrument, and revealed a statistically significant gain in interest over 20 weeks, p = 0.04, as predicted.
Qualitative analyses from field observations, video analysis, and participant interviews, reveal how peer interactions maintained and developed interest as participants moved closer to full participation in the community of practice over the course of the robotics competition. By combining quantitative measurement of interest with longitudinal qualitative analysis of participant interactions, this research contributes to our empirical and theoretical understanding of the emergence, development, and maintenance of interest in after-school settings, with
implications for how to best design such programs in order to broaden participation and engagement in engineering.

The Four-Phase Interest in Engineering Survey (FIDES) was developed to address a need for a surve... more The Four-Phase Interest in Engineering Survey (FIDES) was developed to address a need for a survey instrument derived from current interest development theory to measure the psychological construct of student interest in engineering. The FIDES survey was developed using the Four-Phase Model of Interest Development as a theoretical framework for identifying indicators of student interest. In this paper we discuss the design and development of the FIDES instrument using an iterative method. After two rounds of modification and fine tuning, we present the FIDES 2.0 instrument as a valid and reliable tool for measuring student interest in engineering. We explore the merits and limitations for the FIDES 2.0 to inform our understanding of interest as a developing psychological construct, and suggest appropriate applications that would benefit from using the survey.

Mathematics curricula are frequently rich with visuals, but these visuals are often not designed ... more Mathematics curricula are frequently rich with visuals, but these visuals are often not designed for optimal use of students’ limited cognitive resources. The authors of this study revised the visuals in a mathematics lesson based on instructional design principles. The purpose of this study is to examine the effects of these revised visuals on students’ cognitive load, cognitive processing, learning, and interest. Middle-school students (N = 62) read a lesson on early algebra with original or revised visuals while their eye movements were recorded. Students in the low prior knowledge group had less cognitive load and cognitive processing with the revised lesson than the original lesson. However, the reverse was true for students in the middle prior knowledge group. There were no effects of the revisions on learning. The findings are discussed in the context of the expertise reversal effect as well as the cognitive theory of multimedia learning and cognitive load theory.

Conference Presentations by Joseph Michaelis

Images in mathematics textbooks are intended to be informative, stimulating, and to catch and hol... more Images in mathematics textbooks are intended to be informative, stimulating, and to catch and hold student interest. How images are used in concert with text greatly influences student comprehension and learning (e.g., Butcher, 2006). Instructional design principles based on cognitive processing theories prescribe best practices for presenting images along with text in learning materials (Mayer, 2009). Yet little previous work has examined cognitive processing in widely used, commercially-available, mathematics curriculum materials. Working with the creators of the Connected Mathematics Project (CMP2; Lappan et al., 2006), we made revisions to the commercial materials, based on three principles: Assisting in the integration of information between text and images (Contiguity), highlighting mathematically relevant information (Signaling), and eliminating potentially distracting images (Coherence; Mayer, 2009). Our revisions were reviewed for inter-rater reliability, and approved by mathematics educators and curriculum developers. Participants were 57 middle-school students entering sixth or seventh grade (M = 11.12 years). Eye-tracking measures were used to study cognitive processing based on the eye-mind assumption (Just & Carpenter, 1980), which states that the location of a person's visual attention indicates what they are currently cognitively processing. We compared first pass and second pass dwell times for images and text across Signaling, Contiguity, and Coherence revisions. First pass dwell time was the total duration of fixations in an area of interest from first entry to exit (Holmqvist, et al., 2011), with longer processing time indicating greater cognitive load (Hyönä & Nurminen, 2006). Second pass dwell time was defined as the total duration of all fixations in an area of interest after first exiting the area (Hyönä et al., 2003), and served as an indicator of how much time was spent on integrative processing (Hyönä & Nurminen, 2006). For lesson pages which included Signaling and Contiguity revisions, second pass dwell time on the image was greater for students in the revised condition (F(1,55) = 4.47, p < 0.05). Increased second pass dwell time on images in the revised condition was positively correlated with higher learning outcomes (r = 0.52, p = 0.004). Previous research found that increased second pass dwell times on images are positively associated with learning (Rau, Michaelis, & Fay, 2015). While the cognitive processing of images with signaling and contiguity features is less efficient, the additional dwell time benefits learning. For pages that included revisions where an irrelevant image was removed (Coherence), first pass dwell times on the text were negatively correlated in the original condition (r =-0.42, p = 0.032), but positively correlated in the revised condition (r = 0.380, p = 0.046) with higher post-test scores (Figure 1). For these pages, we found that higher post-test scores were positively correlated for both second pass dwell times on the image in the original condition (r = .47, p = 0.013), and on the text in the revised condition (r = 0.45, p = 0.016). We contribute to the theory of cognitively based design principles by investigating when greater processing demands are desirable for increased learning.