"Evaluating the Effectiveness of Tablets and Smartphones for Learning Data Science: Opportunities and Limitations"
Description
Descriptive (example) Overall Pretest mean ≈ 50 ± 10. Overall Posttest mean ≈ 62 ± 11. Average Gain by device (approximate): Tablet: mean gain ≈ 13–15 points Laptop: mean gain ≈ 11–13 points Smartphone: mean gain ≈ 9–11 points ANOVA F-statistic: expect F > 5, p < 0.01 (indicating at least one group differs). Tablet vs Smartphone (t-test) t statistic: positive, p < 0.05 (tablets better on average than smartphones). Regression After adjusting for Hours, Motivation, Usability, and prior experience, tablet device dummy retains a positive coefficient vs smartphone; magnitude depends on data noise but typically ~ +1.5–3 gain points. Dropout Dropout rate primarily correlated with poor internet and low motivation. Expect dropout around 8–15% overall depending on the synthetic parameters.
Files
Steps to reproduce
Reproducibility & Code The full Python script above reproduces the dataset and analysis. To produce a publication-quality appendix, run the script and export the CSV; then use the CSV for tables and figures (Excel, R, or Python plotting). If you want, I can generate polished figures (boxplots, violin plots, model coefficient plots) — tell me which file location you used and I’ll provide plotting code (or convert to R with ggplot2).
Institutions
- Moorfields Private