Summary:
- The article discusses the pedagogical challenges in statistics and data science education, specifically the "clean data/clean model" trap where students are shielded from the complexities of real-world data processing.
- It advocates for teaching students how to handle messy, incomplete, and noisy data to better prepare them for authentic scientific research and statistical modeling.
- The author emphasizes that model building is an iterative, diagnostic process rather than a linear execution, highlighting the importance of teaching students to identify and resolve discrepancies between their models and the empirical data.