A new pedagogy to address the unacknowledged failure of American secondary CS ed

Key Takeaways from Scott Portnoff (Author) from the SIGCSE Mailing List:

Part 1 argues that the survey courses ECS and APCSP have questionable, if any, value in terms of content or pedagogy, and have perpetuated an inequitable 2-tier system in large urban school districts like Los Angeles USD. District data for the 2018 and 2019 APCSP exams reveals overwhelming failure rates for under-represented minority (URM) students, precisely the demographic groups the course was meant to benefit.

Part 2 discusses a pre-APCSA programming course that incorporates instructional strategies to address a crucial and missing pedagogic piece - the teaching of programming languages as languages per se - in order to counter the effects of the decades-old Novice Programmer Failure Problem. Cognitive evidence supporting this approach now include 3 fMRI studies concluding that understanding of programming languages occurs in the language processing regions of the brain, not math, not logic. If scaled, this pedagogy could better prepare all students, but particularly URMs, for successful completion of the APCSA course and exam, and so boost successful participation rates in college.

Part 3 showcases a concrete example of this approach using Codingbat. Grammar rules can be thought of as generalizations of language patterns. Instructional strategies therefore focus on providing repetitive and incrementally more complex practice with (a) Patterns to clarify how API methods, control structures, parameters, variables, and like syntactic features themselves are used, and (b) Patterns that combine these features in crafting solutions to specific types of beginner problems, prioritizing coding paradigms that are clear and simple.