The importance of data science skills for modern scientific research cannot be understated. Although policy documents increasingly recommend what skills should be included in undergraduate statistics and data science curricula, little is known about how students actually develop and apply these skills. This paper addresses this gap through an in-depth case study tracing one student’s learning progressions throughout her master’s program. Using a qualitative method to analyze student code, which has seen little use in statistics education research, I examined how Alicia transferred the data science skills from her applied statistics course into authentic research settings. The analysis shows that, while Alicia successfully navigated new challenges, she encountered persistent hurdles when extending bivariate techniques into multivariate contexts, particularly with visualizations and summary statistics. These findings highlight the obstacles students may face when applying classroom knowledge to real-world data problems. The results carry implications for instructors designing curricula, researchers studying how students learn data science, and policymakers shaping educational standards, underscoring the need to pair policy recommendations with research on the realities of student learning.

R Script files submitted by Alicia (pseudonym) over the course of the study. The files are named according to when they were submitted:

• December 2018
  • R Script #1
• April 2019
  • R Script #1 (revised)
  • R Script #2
• September 2019
  • R Script #1 (revised)
  • R Script #2 (revised)

Qualitative Data Analysis Files (Rich text files)

• December 2018 Script #1
• April 2019 Script #1
• April 2019 Script #2
• September 2019 Script #1
• September 2019 Script #2

Quantitative Data Analysis Files

• r-code-themes.csv
  • Comma separated values file with separate sheets for each R script
  • Each sheet contains the qualitative code assigned to each line of code and whether the code contained errors.