United States

Earning a bachelor&#39;s degree is expensive and time-consuming. Many undergraduate students pursue Advanced Placement (AP) courses in high school or transfer coursework from other degree-granting institutions. However, the effect of those transfer courses on the time it takes students to graduate (time-to-degree) is currently not well understood. In this work, we investigated how incoming transfer courses impact students’ time-to-degree. To do so, we explored how well a logistic regression model would predict whether a student graduated before or after 8 semesters, 10 semesters, and 12 semesters. When discerning time-to-degree, there are several other features that were also considered in addition to transfer status: students’ demographic information such as race/ethnicity, gender, first-generation status, and their academic background preparation as measured by their high school GPA and math placement scores. Here, we will discuss the findings of our model as well as the effects of data imputation on our results.

AAPT SM 21

Investigating Impact of Transfer Credits on Time-to-Degree using Regression Models

time-to-degree

regression

transfer

The 2020 AAPT Virtual Summer Meeting brought together physics educators from around the globe to learn about the latest in physics theory, research methodology and education from leaders in their field. The program featured plenary speakers, sessions, workshops and many special activities.

For more information please visit the 2020 AAPT Virtual Summer Meeting page at: https://www.aapt.org/Conferences/SM2020/index.cfm.


AAPT SM 20

The 2020 AAPT Virtual Summer Meeting brought together physics educators from around the globe to learn about the latest in physics theory, research methodology and education from leaders in their field. 

Earning a bachelor's degree is expensive and time-consuming. Many undergraduate students pursue Advanced Placement (AP) courses in high school or transfer coursework from other degree-granting institutions. However, the effect of those transfer courses on the time it takes students to graduate (time-to-degree) is currently not well understood. In this work, we investigated how incoming transfer courses impact students’ time-to-degree. To do so, we explored how well a logistic regression model would predict whether a student graduated before or after 8 semesters, 10 semesters, and 12 semesters. When discerning time-to-degree, there are several other features that were also considered in addition to transfer status: students’ demographic information such as race/ethnicity, gender, first-generation status, and their academic background preparation as measured by their high school GPA and math placement scores. Here, we will discuss the findings of our model as well as the effects of data imputation on our results.

technical paper

Join AAPT and Physics Educators worldwide in developing and strengthening your professional connections by attending the 2021 AAPT Virtual Summer Meeting. AAPT meetings are a great way to reflect on your teaching, recommit to best practices and re-energize your career with new research, products, and techniques.

Listen to internationally known plenary speakers, attend sessions, workshops, and special activities. Attending the virtual summer meeting is an excellent opportunity to share your expertise and learn from and network with leaders in the field.

Please log in with your AAPT account or register for the event if you haven't already.

If you are registered please note that you need to be logged in with your AAPT email / credentials in order to get access.

The Colorado Learning Attitudes about Science Survey (CLASS) assesses students’ perceptions of physics through self-reported agreement with attitudinal statements. Previous work shows that CLASS results may be subject to response-shift bias, in which knowledge gained during an intervention causes learners’ perceptions of their pre-instructional state to become more novice-like. We explored possible sources of this response-shift bias in the CLASS by asking students to (1) complete the survey twice at the beginning of the semester (providing their pre-instruction responses and expected end-of-semester responses) and twice at the end of the semester (providing their post-instruction responses and retrospective beginning-of-semester responses), and (2) comment on their concurrent responses. By comparing beginning-of-semester and end-of-semester comments with survey shifts from a broad sample of students, we illuminate an underexplored dimension of student thinking. Preliminary results show that our students expect their perceptions of the applicability of physics will grow directly from acquiring physics knowledge.

Sources of Response-Shift Bias in the CLASS’s Real-World Connection

Educators and researchers often use research-based assessments to measure the efficacy of their courses. Limited resources exist, however, for interpreting assessment results. We will present analyses and representations created with data from more than 500 introductory physics courses that provide a context for interpreting assessment results. The data set comes from the online Learning About STEM Student Outcomes (LASSO) platform and the scientific literature. The representations include scatterplots of pretest and posttest scores and distributions of effect sizes, Cohen’s d with Hedges’ correction. These representations show how courses compare to the larger database before and after instruction. The results show that common rules of thumb for effect sizes are not useful for interpreting effect sizes on research-based assessments. Excel and R files are available at https://tinyurl.com/stemequityresources for readers to create their own visualizations overlaying their course data on the larger data set.

Tools for Identifying Effective Courses

As a method to shorten the test time of the Force Concept Inventory (FCI), we suggest the use of Computerized Adaptive Testing (CAT). CAT is the process of administering a test on a computer, with items (i.e., questions) selected based upon the responses of the examinee to prior items. In so doing, the test length can be significantly shortened. As a step to develop a CAT–based version of the FCI (FCI-CAT), we examined the optimal test length of the FCI-CAT such that accuracy and precision ([which we measure in terms of root-mean-square error (RMSE)] of Cohen’s d would be comparable to that of the full FCI for a given class size. We conducted a Monte Carlo simulation to analyze how the RMSE of Cohen’s d depend upon the test-length. As a result, we found, that for a class size of 40, we may reduce the test length of the FCI-CAT to 15-19 items, thereby reducing the test time of the FCI to 50% - 63%, with an accompanying decrease in accuracy and precision of only 5-10%.

Optimizing the length of computerized adaptive testing for the FCI

The presented study investigates the use of mathematical procedures to solve items in a kinematic context. For this purpose, a validated test instrument by Ceuppens et al. (2019) was administered to high school students which included pairs of items in mathematics and kinematics contexts that are isomorphic to each other, i.e., have the same surface features. 
While solving the tasks, eye movements of N=35 learners were recorded by a stationary eyetracker. For selected items, we demonstrate that gaze behavior and thus visual strategies differ between solving the mathematical and kinematic items. 
The results show that specific solution strategies are induced by the task context, which means that procedures that can be applied in the mathematical context cannot be transferred to the kinematic context. The talk presents the study and discuss the results in details.


Gaze data analysis in the interpretation of linear graphs

Assessment scores can be used to inform instructors about their students’ learning, ideally to inform subsequent course modifications to better facilitate students’ learning. However, explicit course modifications require interpretations about what assessment scores convey instructors about the extent to which students meet the learning goals instructors have for their students, i.e., actionable feedback for instructors (hereinafter feedback). As part of developing a standardized assessment for upper-division thermal physics, we articulate a theory-based approach to generating feedback that goes beyond simply reporting students’ scores. The extent to which students met the learning goals instructors have for them guides our feedback generation. In this talk/poster, we present our theory-based approach to feedback design by integrating effective feedback practices that promote instructors’ agency. This work emphasizes the development of feedback for instructors to better facilitate students’ learning in physics classrooms. 


Research-based Assessment Feedback for Instructors

Can roles improve group cooperation and make group-work more equitable? During the past year, we developed a series of hour-long collaborative activities to supplement individual investigations for our introductory physics lab course. Students worked in groups of 4 in Zoom breakout rooms with the iOLab lesson player. We developed the activities to revolve around 4 student roles: manager, archivist, experimentalist, and skeptic. Analysis based on student work, interviews, and survey responses suggested that the roles helped support student collaboration, especially when the roles where integrated into the tasks we asked students to complete. However, by themselves, group roles were not sufficient to prevent negative gendered group interactions.

Roles in Collaborative Introductory Lab Activities

Modified Module Analysis using partial correlations (MMA-P) has proven to be a productive method for studying conceptual inventories. MMA-P can be used to analyze large datasets by forming networks of responses then using community detection algorithms to identify sets of responses that are consistently selected together. MMA-P can be extended by using multilevel network analysis. A multilevel network was constructed with pretest and post-test Force Concept Inventory (FCI) responses from five different institutions using samples ranging from N=10,000 to N=200. Each layer of the multilevel network was represented by a network made from the application of MMA-P to data from each institution. Within each layer, MMA-P finds communities of popular responses. Network comparison methods were then used to understand the similarities and differences between the sets of nodes and communities in the networks. This provides insight into the generality of previously identified properties of the FCI between institutions.

Multilevel Module Analysis of the FCI

This talk examines the retention of physics majors at a moderately sized physics department which struggles with increasing the number of graduates using 20 years of graduation data. Survival analysis identifies the point of highest risk for two potential paths out of the major: leaving college and changing majors. Substantial risk of losing majors exists through the first two years of college. Logistic regression is used to explore the factors related to retention in general and through the first two years of college. Decision trees, a method applied in machine learning, is used to quantify the thresholds of these variables important to evaluating risk.


Exploring Factors Influencing the Retention of Physics Majors

Research on self-efficacy -- or the confidence in one’s capability to perform a task -- has been shown to predict achievement and persistence in the sciences. Throughout a student’s experience in curricular and/or co-curricular activities, they may face threats to their self-efficacy. To investigate the threats or supports to students’ self-efficacy, we are using a mixed methods approach. This novel design creates a holistic narrative of which curricular and/or co-curricular activities are impacting a student’s self-efficacy as well as how and why these activities may be affecting their self-efficacy. We use a mixed methods approach to collect in-the-moment information about lived experiences. Quantitatively, we employ the Experience Sampling Method (ESM) using surveys of domain-specific self-efficacy. Qualitatively, we use daily reflections to investigate threats and supports toward students’ self-efficacy. Using an example, we will present the research design while focusing on building this conversation between the ESM and qualitative reflections.

Identifying Threats and Supports to Self-Efficacy Using Mixed Methods Research

Modeling is an important skill in experimental physics and thus a learning outcome of many laboratory classes. One promising approach to encourage modeling is to incorporate multiweek open-ended projects within lab courses. However, not much work has been done on assessing student engagement with modeling in actual courses. We present an analysis of projects from three upper-level university lab courses at different institutions where we investigate how the students engage with modeling and how this engagement differs depending on features of the projects. The projects in our dataset vary widely with different degrees of model construction. We observe correlations between features of the projects and the ways the students engage with modeling, which may have implications for how instructors guide students towards choosing certain types of projects.

Premium content

Downloads

Next from AAPT SM 21

Sources of Response-Shift Bias in the CLASS’s Real-World Connection

Similar lecture

Analyzing Time-to-Degree for Transfer Students at Michigan State University