Better Data, Better Progress: Methods for Measuring Inequities in Access to Computer Science Education

• In Event: 47.027 - CS for All: An Intersectional Approach to Unpacking Equity in Computer Science Education

Abstract

This study enhances research on CS education for underrepresented populations by collecting and publishing more accurate and comprehensive state-level data than what has previously been available in the field. Equity here refers to both access to and successful completion of high school CS courses by traditionally underrepresented populations such as students of color, females, and students from low income families. Research on equitable access to CS education has been hindered by the fact that available data have tended to be lacking in either precision or scope. For example, College Board (2018) measures of student enrollment in CS by different demographics only address two Advanced Placement CS courses. The Google Inc. & Gallup Inc. (2016) report provides estimates of the percentage of schools that offer any CS course, but the estimates come from self-reported data that inherently lack precision and come with other limitations, including concerns of construct validity. For example, across all sampled schools in the U.S., 78% of high school principals reported that their school offers at least one CS course. Numbers like these are likely to be inflated as educational leaders often conflate computing (e.g., programming) with computer literacy courses (e.g., keyboarding). Indeed, our research using objective counts shows that only 39% of high schools in Texas offer at least one CS course.

Ensuring equitable participation in CS education requires reliable and systematic methods for tracking key outcome measures longitudinally. The limitations of the extant research and available data underscore the need for data on CS course offerings and enrollment that come from systematic and objective measurement methods. Such methods are possible at the state level, as many states already have a mandatory data reporting system in place for public schools. This study took advantage of one such data reporting system in Texas by using actual counts from every school in the state to analyze demographic enrollment trends and determine the exact nature of the gap in access to CS education for underrepresented students. Results include student-level data that illustrate how enrollment gaps in terms of gender, income, and race differ across 16 distinct CS courses and how these gaps are exacerbated for students who live in rural communities. Results from school-level data demonstrate how the number of CS courses that schools offer differs according to its location and the demographic makeup of its student body. In conjunction with the school-level data, we present the CS Equity Quotient (CSEQ), a new method for quantifying the degree to which a campus or district’s CS course enrollment reflects the gender, income and ethnic diversity of the school. It is our intention that the CSEQ serve as a consistent, simple metric to assess CS equity and a model for other researchers nationally to accurately measure and track broadening participation in CS efforts even as overall student demographics may change over time.

Authors

• Jayce R. Warner, The University of Texas at Austin

• Carol Fletcher, The University of Texas at Austin

• Lisa Garbrecht, The University of Texas at Austin