PERC 2016 Abstract Detail Page
Previous Page | New Search | Browse All
| Abstract Title: | Expanding Research Questions by Expanding Quantitative Methodologies |
|---|---|
| Abstract: | This poster symposium will highlight multiple quantitative methodologies that can allow physics education researchers to expand the type of research questions we can address by exploring the alignment between research design, research questions, data analysis and knowledge claims. As our guiding theoretical frameworks push us to examine our participants in richer detail, we must simultaneously adapt our research design and analysis methods to account for these multi-faceted data. Presenters in this session will highlight four distinct methodologies, requiring advanced quantitative analysis techniques, that allowed them to develop detailed quantitative descriptions of students' learning and attitudes. By comparing across presenters, participants will be able to trace how differences in each researchers' guiding theoretical framework led to specific decisions about research design and analysis, allowing them to answer different types of research questions. |
| Abstract Type: | Poster Symposium |
Author/Organizer Information | |
| Primary Contact: |
Jacquelyn Chini University of Central Florida 4111 Libra Dr. Orlando, FL 32816 Phone: 4078233607 |
Symposium Specific Information | |
| Moderator: | Jacquelyn Chini |
| Presentation 1 Title: | Finding groups in student-level data: utilizing the Profile Approach |
| Presentation 1 Authors: | Jarrad W. T. Pond and Jacquelyn J. Chini |
| Presentation 1 Abstract: | Individual students bring their own motivations and self-regulated study strategies to the classroom. When interested in characterizing students, one can turn to the Profile Approach. At the core of this approach is a theoretical framework that considers the interactions of different motivational and strategic self-regulatory constructs and describes the various, distinct patterns that exist across these constructs as Learning Profiles. The Profile Approach relies on individuals' characteristics; thus, accompanying research designs must incorporate collecting student-level data on various motivational and strategic self-regulatory constructs. Also, analyses must consider patterns across these constructs. Cluster analysis is such a technique, allowing identification of coherent groups (Learning Profiles) based on patterns in the data. Resulting Profiles are useful for exploring hypotheses about student characteristics, guiding instructors to better understand their students, and follow-up statistical analyses. I present--from start to finish--applying the Profile Approach to identify Learning Profiles among algebra-based, studio-mode physics students. |
| Presentation 2 Title: | Using structural equation modeling to test the physics identity framework |
| Presentation 2 Authors: | Robynne M. Lock, Zahra Hazari, Geoff Potvin, and Jennifer Cribbs |
| Presentation 2 Abstract: | Fully modeling the complex relationships between multiple factors in a theoretical framework requires connecting several statistical techniques. Structural equation modeling (SEM) combines confirmatory factor analysis (CFA), path analysis, and multiple regression. Rather than providing a form of exploratory data analysis, SEM is a confirmatory technique that involves testing and modifying a hypothesized model. Specific tests indicate ways to improve the model, and a variety of fit indices are used to assess the model fit. The implementation of SEM will be shown through the example of testing the physics identity theoretical framework. This includes performing the initial CFA to verify the identity dimensions and assessing model fit. Additionally, the method of modeling multiple groups will be demonstrated by comparing the physics identity model for males and for females. Examples of implementation will be shown in R. Furthermore, the types of research questions to which SEM is most applicable will be described. |
| Presentation 3 Title: | Addressing Relational Data in Students' Representation Use with Network Analysis |
| Presentation 3 Authors: | Daryl McPadden, Jesper Bruun, Eric Brewe |
| Presentation 3 Abstract: | Network analysis is aligned with a theoretical framework that values relational questions and connection between quantities. One of the problems of using frequentist statistics in education is that most statistical analyses (such as t-tests and analysis of variance) require normally distributed data set, with observations that are independent of one another; however, when studying students in a classroom, these assumptions are often violated. Students interact with each other, instructors, and course material, allowing ideas to be transferred in a variety of ways. Learning is embedded in these interactions and can be an exciting area to research but fundamentally requires a research design that captures relational data. This poster will present an example from our work on students' use of representations to demonstrate the types of questions that network analysis can answer, highlight a variety of the analyses that can be run with relational data, and address some of the challenges we faced in using this methodology. |
| Presentation 4 Title: | A Multi-faceted Approach to Measuring Student Understanding |
| Presentation 4 Authors: | Trevor I. Smith, Ian T. Griffin, and Nicholas J. Wright |
| Presentation 4 Abstract: | Data from the FMCE support findings that students' understanding of graphs impacts their abilities to express their understanding of the relationships between and among forces and various quantities of motion. We approach the current study with an assumption that measurements of student understanding of a particular topic depend not only on the student but also on the instrument used to make the measurement. Multiple measurements are needed to build a more complete picture of what the student actually "knows." We compare individual student responses to 12 specific questions on the FMCE from three question clusters. Each cluster provides students with four identical descriptions of object velocities. In one cluster students choose a description of an accompanying force, in the others students choose a graph of force or acceleration vs. time. We use Cohen's w to report the correlation between the clusters and consistency plots to show the impact of instruction. |
