Generative AI presents many possible use cases in the context of education research, including summarizing previous research, creating literature reviews, generating synthetic data, analyzing qualitative as well as quantitative data, creating data visualizations, and writing research papers. However, there are also many ethical issues that may arise related to the use of generative AI, including concerns related to environmental impact, human impact, data privacy, transparency and replicability, the potential for data re-identification, intellectual property and intellectual debt, the digital divide, and accuracy and bias. In order to promote the responsible use of AI in education research, we formed a group with expertise across the relevant research practices and ethical issues to convene in order to articulate guidelines for the use of generative AI in STEM education research.

While the U.S. government tracks teacher certification, this tracking is limited to initial certification. However, many computer science (CS) teachers did not initially certify in CS but in another subject. As a result, there is a data gap involving pathways to CS teacher certification, pathway requirements, and the number and characteristics of teachers who pursue each pathway. This gap in the data landscape may be an impediment to the goal of expanding CS education: a lack of understanding of these pathways may inhibit recruitment, retention, and ensuring that the pathways adequately prepare educators to support student learning.

The arrival of generative AI in schools has altered the education landscape in myriad ways and raises issues from the pedagogical to the ethical. In this context, the perceptions of teachers regarding whether, what, and how students should be taught about AI are important to consider. For this study, we conducted semi-structured interviews with several dozen teachers, representing a wide variety of teaching contexts. Our analysis of the interviews revealed that most teachers support student learning about AI, but that their reasoning varied substantially. Rationales for supporting education around AI included encouraging best practices for AI use, preparing students for their futures, equipping students to critically evaluate AI output, and promoting academic integrity, among others. Other teachers were less supportive of AI education, worried that encouraging its use would engender over-reliance on it. Teachers also expressed a variety of perspectives on the ways in which AI might either help or harm learning, augmenting concerns about what and how students are taught about these tools. These educator perspectives constitute an important contribution toward the wider conversation about AI in schools.

In the United States, state learning standards guide curriculum, assessment, teacher certification, and other key drivers of the student learning experience. Investigating standards allows us to answer a lot of big questions about the field of K-12 computer science (CS) education. Our team has created a dataset of state-level K-12 CS standards for all US states that currently have such standards (n = 42). This dataset was created by CS subject matter experts, who - for each of the approximately 10,000 state CS standards - manually tagged its assigned grade level/band, category/topic, and, if applicable, which CSTA standard it is identical or similar to. We also determined the standards' cognitive complexity using Bloom's Revised Taxonomy. Using the dataset, we were able to analyze each state's CS standards using a variety of metrics and approaches. To our knowledge, this is the first comprehensive, publicly available dataset of state CS standards that includes the factors mentioned previously. We believe that this dataset will be useful to other CS education researchers, including those who want to better understand the state and national landscape of K-12 CS education in the US, the characteristics of CS learning standards, the coverage of particular CS topics (e.g., cybersecurity, AI), and many other topics. In this lightning talk, we will introduce the dataset's features as well as some tools that we have developed (e.g., to determine a standard's Bloom's level) that may be useful to others who use the dataset.

Promoting equity in computer science is crucial, but how best to measure equity is not entirely clear. Building on the tradition of critical quantitative inquiry, we present a test case of four different ways to measure equity in undergraduate CS programs for a given target population: the count of the population, the percent of the population in CS (as opposed to in other majors), the percent of CS students from the target population, and a metric we introduce, the proportional representation index (PRI). The PRI mitigates the impact that very large schools, very large CS programs, and regional population variation can have on the other metrics, but it also presents its own challenges. We believe that this work will contribute to the conversation regarding measuring equity by presenting and then problematizing several possible metrics.

In recent years, eight states have adopted a graduation requirement in computer science (CS), and other states are considering similar requirements. Due to the recency of these requirements, little is known about student and teacher perceptions of course(s) that fulfill the requirement and their content. This project seeks to answer the question, What are the perceptions of students who are studying CS beyond high school and CS teachers of a high school CS requirement and its content?

This full research paper explores ways in which student parents who are studying engineering do and do not differ from other students who study engineering. INTRODUCTION: About one-fifth of undergraduate students are parents, and these students are likely to differ somewhat from other students. However, little research has been conducted on student parents, particularly those who are studying engineering. Further, engineering continues to suffer from disproportionate representation of students from historically marginalized gender and/or racial/ethnic backgrounds, and the equity implications of parental status are currently poorly understood. OBJECTIVE: The purpose of this paper is to answer the research question, How do engineering students who are parents differ from other engineering students? METHODS: We analyzed data from the U.S. 2020 National Postsecondary Student Aid Study for undergraduate students, which indicates which students have dependent children. In this exploratory data analysis, we compared engineering students who are parents to those who are not across a variety of demographic, financial, and academic characteristics. RESULTS: We found that engineering students are far less likely than other students to be parents (6% v. 18%). Engineering students who are parents differ from other engineering students in that they are more likely to be older, to be men, to be Black or African American, to have a job, to attend school part-time, and to attend a nonselective college or a two-year college. They are also more likely to have no funding for tuition other than a credit card and to experience food insecurity. They are, surprisingly, less likely to attend a school that offers child care. However, engineering students do not significantly differ with respect to their status as parents with regard to disability status or in their likelihood of having a high or low GPA. DISCUSSION: Our findings suggest that engineering students who are parents face a distinct set of challenges (e.g., financial insecurity) but also show some strengths (e.g., similar GPAs) despite the many demands on their time. Further, the preponderance of Black engineering students who are parents suggests that better meeting the needs of student parents may be one vector for improving representation in engineering education.

Introduction: This full research paper explores how data is used to design equity interventions. The percent of engineering degrees awarded to people who have historically been excluded from engineering has increased since 2010. But there is still substantial underrepresentation for women and people from racially and ethnically marginalized groups. One potentially promising practice is to harness the power of data, including big data, to design interventions to improve equity. However, there are gaps in our understanding of how higher education faculty and staff use this data in designing equity interventions. Objective: The purpose of this paper is to answer the question: How is data used in the process of designing efforts to improve equity in engineering programs? Methods: We performed a content analysis on surveys completed by and artifacts generated by higher education faculty and staff who participated in a structured professional development and research experience. The experience focused on planning and executing a data-driven project designed to improve equity in engineering education. Results: Analysis of the data suggests the following: participants articulate the core challenge that they are facing in terms of data indicating demographic disparities, make the case for addressing inequities via presentation of relevant data, and conceptualize evaluating success via gathering quantitative data about intervention outcomes. However, many projects are not, themselves, focused on the creation or use of data (e.g., mentoring programs), Conclusion: This work shows how higher education faculty and staff who are designing interventions to promote equity in engineering education are using data in the design phase of their work. Understanding their patterns of data use is an important first step in determining how data use impacts their projects' outcomes and, based on that finding, how future cohorts can be best supported in the design of their interventions.

Improving equity in K-12 Computer Science (CS) education benefits from the collaboration of classroom teachers, school counselors, and school leaders. This paper presents the outcomes of a pilot program that brought together cross-functional teams consisting of CS teachers, school counselors, and administrators. Over the course of a year, these teams attended monthly, equity-focused workshops, leveraging pre-existing materials from affordable, high-quality, research-based programs. The use of these resources demonstrated benefits of sequencing and synthesizing existing programs. Evidence from surveys and interviews shows that the workshops promoted learning and fostered collaboration between the cross-functional teams that would not have happened otherwise. Participants were motivated by the program, and they generated ideas that turned into actionable projects to promote CS education equity in their schools. While the initiative was well received, areas for improvement were identified, particularly, in school recruitment, workshop structure, and evaluation. This pilot initiative demonstrates that equity-centered programs comprised of cross-functional teams can help achieve systemic improvement of CS education equity.

Correcting the under-representation of people with disabilities in computing and computing-related careers will require a research base that enables equitable computing education (CE) for all students. For this study, we analyzed research papers in order to better understand the landscape of CE research as it relates to students with disabilities. Of the 771 papers reviewed from 12 major CE venues over the last decade, very few (n = 14) specify student participants' disability status; this <3% rate is under-representative, given that about 15% of students in the US have a disability. Studies that do specify disability status tend to involve students who are in the middle grades, in the United States, and/or are participating in elective activities. Demographic factors (e.g., gender and race) are specified about half of the time and analyzed about half as often as they are specified. A small cadre of researchers (51 from 22 different institutions) are responsible for the studies specifying disability status. Studies - whether they specify disability status or not - most commonly use block-based programming, which may present accessibility issues for some students. Advancing equity for all students will necessitate much more research on the computing experiences of students with disabilities, including reporting factors such as demographic data that make it easier to determine whether and how the research applies in a given context. More research focused on accessible tools and languages for learning computing is also needed to further understand promising practices for teaching students with disabilities.

Traditionally, computer science (CS) in the United States has been an elective subject at the high school level. In recent years, however, some school systems have created a CS graduation requirement. Designing a required CS course that meets the needs of anticipated future advancements in the field necessitates exploring the research question, What computing content do high school teachers, college instructors, and computing industry professionals prioritize in a required computer science course for high school students? To better understand what these different groups perceive to be the essential content of a foundational high school CS course, we conducted a series of focus groups. These focus groups explored participants' (n = 21) thinking about what content would be most important to prioritize in a required high school CS course. Transcripts of the focus groups were abductively coded and then analyzed to determine what CS content priorities were identified and what disagreements about priorities exist.

Problem. This full research paper describes the results of a literature review of data collected about K-12 computer science education research and initiatives. Over the last ten years, K-12 computer science (CS) education research has evolved to meet the needs and progress of the computer science education community. As a field, however, we have no empirical evidence of what this evolution is and how it has managed to produce the empirical evidence needed to support the long-term goals of computing education research as computing education grows into every primary and secondary classroom. Research Question. Our question for this study was How has K-12 CS education research evolved over the last decade, including when examining the research based on standards and inclusion of student participants? Methodology. Using a publicly available database of curated articles documenting K-12 computing education research efforts, queries were run to extract pre-specified subsets of data. Descriptive statistics were calculated to identify trends to support answers to the research question. Findings. We consider how this data reflects on changes in the last decade in light of the CS for All initiatives, including the fact that less than 5% of studies include students with disabilities and less than 20% of studies report participants' socioeconomic status. With the growth of computing in schools in the last decade, it is important to consider previous computing experience of students when analyzing results, but our data shows that only 40% of studies do. Implications. The good news is that this increase means that more topics, contexts, and student subgroups can be explored each year, and new insights on best practices are likely to arise from these findings. The bad news is that the attention to the needs of historically marginalized groups is not increasing at the same pace. While our data shows modest to gradual increases in the percent of research articles that report on factors such as the student race and instructor gender, these factors are still rarely reported, which may make it more difficult to ensure that CS education research involves a wide diversity of participants.

Equitable instruction requires that students who are part of a linguistic minority (i.e., they use a language different from that of the majority in their area) are provided instruction that allows them to accomplish course goals. To date, studies on the experience of novice programmers who are in a linguistic minority are quite rare. This poster is an effort to address that gap: it examines the programming process data of students learning Java (n = 2,621) from around the world, some of whom were a linguistic minority. Results indicate that there are statistically significant differences between students in linguistic minorities and majorities for two types of interaction with the IDE: editing events and file events. This finding is one step toward understanding the programming behavior of linguistic minorities.

A variety of approaches have been used in computer science education research; one relatively new addition is the use of very large data sets. This dissertation will use one of these data sets in order to investigate the process by which novice students learn to program. Specifically, an iterative approach will be used to determine which learning theories are (or are not) supported by the data. The focus will be on which patterns of behavior do (or do not) lead a student to learning a programming concept.

Direct instruction is a learning theory that garners strong support as well as vociferous criticism. This poster analyzes how DI is used in CS education research, finding that most references to DI are brief and usually contrast DI to another learning theory. In cases where DI is part of a study's research design, it is usually the case that an intervention led to better learning outcomes than DI. Additionally, many studies note a student preference for more DI.

This dissertation will use the Blackbox data set to explore which student behaviors are most likely to lead to learning a programming concept, resulting in a model of student learning which will be analyzed to determine which learning theories and models it supports. Finally, whether machine learning can be used to predict student learning will be explored.

Against a backdrop of relatively high failure rates in introductory programming courses and a persistent lack of diversity, computer science education (or CSEd) still has substantial room for improvement. Central to that effort is determining which theories of learning are most appropriate to CSEd and how best to apply them. One prominent theory -- constructivism (hereafter C*)-- is frequently referenced in the research literature. This ongoing project explores the use of C* in CSEd research using citation analysis of journal articles and proceedings published in the wake of Ben-Ari's foundational work on the topic. Using a novel Python tool to parse BibTex data, this project explores trends in the use of C* in CSEd research. There were 536 articles in the data set. All articles included C* in their text, with 9% including C* as a keyword and 3% including it in the title. Preliminary findings include the observation that the use of C* has been roughly stable over time and tends to be linked to topics such as pedagogy, active learning, and visualization. Future work could incorporate the text of the article itself into the analysis to achieve a more finely grained view of how C* functions in CSEd research. Another useful data set for analysis would be the hundreds of articles that cite Ben-Ari's paper, especially in light of his recent statement that "constructivism is either nonsense or trivial."

Mathematician Eugenia Cheng offers a fresh framework for thinking about the problem of the under-representation of women in STEM. Cheng’s approach is grounded in her research field, category theory, which leads to her proposal that instead of focusing on a gender binary, we instead consider each person as being on a continuum from “congressive” (that is, focused on interdependence) to “ingressive” (focused on independence). The purpose of this poster is to consider the application of Cheng’s framework to computing education by focusing on three main questions: If we base computing education research on this framework, what might that look like? What might we gain, and what might we lose? Cheng’s framework applied to computing education research would change the methodology used in interventions that consider the participants’ gender. Gains from Cheng’s framework therefore include a way to research, analyze, and implement interventions in computer science education to improve representation that do not reify the gender binary and (further) marginalize non-cis-gendered students. However, adopting Cheng’s framework would create several hurdles, including the need for a valid instrument to assess placement on the continuum. And the framework may well fail to rectify the problem of under-representation. Nonetheless, Cheng’s framework, particularly in its ability to include all students, regardless of gender identity, is worth considering as a tool in computing education, not the least because it suggests a possible contrast to the status quo.

As undergraduate enrollment in computer science (CS) courses continues to expand, concerns surrounding the supply of instructors and the quality of instruction become even more salient. Similarly, computing education and computational thinking programs are expanding in K12 schools but are hampered by concerns over teacher training and curriculum efficacy. This project sought to answer a question: is there a need for a more robust system of networking and resources for computer science education (CSEd) graduate students' The large response to initial community-building efforts indicates that this is likely the case. In just over one year, a global community of 190 graduate students and 90 CS faculty researchers and advisors have registered for a CSEd graduate focused community. Participant funding, while originally intended for in-person conference attendance and face-to-face meetings, has enabled this project to create study groups attended by 73 students and a virtual conference with 101 participants. These activities were established as a result of a needs assessment survey conducted in early 2020. This poster documents the rapid growth of this community and the need students feel for systematic support.

The substantial decline in female enrollment in computer science programs has been a cause of concern, especially as the application of computer technology to an ever-wider range of human endeavor expands. While many studies have examined the attitudes and perceptions of female computer science students, little attention has been paid to studying empirically how classroom interventions-such as curricular choices and pedagogical approaches-may result in disparate impacts in the learning outcomes of male and female students. This study examines recent research articles in leading journals and conference proceedings of computer science education to analyze the gender composition of the subject pool, the extent to which cognitive learning results are disaggregated by gender, and whether there are gender differences in outcomes. This study found that about one-third of studies indicated the gender composition of the subjects, and about one-third of those research subjects were female. About ten percent of papers disaggregated research findings by gender; of those, 30% found differences between male and female subjects. No studies recommended gender disaggregation of data in future work. The lack of data for learning outcomes by gender may be one of the contributors to the paucity of female computer science students; it is possible that interventions found to be more (or less) effective would show different outcomes if the data were considered in light of the gender of the participants. Thus, further research into potential gender differences in the effectiveness of computer science instructional tools is recommended.