Machine learning on pre-pandemic platform click data explains up to 77% of the variance in academic resilience¶

Claim · All claims
Evidence · 1 study · associational r2 · q2
Currency · newest study 2024 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? An elastic regression model including click data pre-processed into 400 categories explains 77% of the variance in resilience, orthogonalized against the behavioral grit measure. → Chen 2024

Evidence¶

Chen 2024¶

Chen, D. L., Ertac, S., Evgeniou, T., Miao, X., Nadaf, A., & Yilmaz, E. (2024). Grit and academic resilience during the COVID-19 pandemic. npj Science of Learning. https://doi.org/10.1038/s41539-024-00265-3

q2 · i? · associational · r2

Elastic regression on platform click data (400 categories, PCA-reduced, N = 1434 in Table 5) yields R² = 0.77, with behavioral grit and selected principal components significant. "77% of variance in resilience is explained by behavioral data measured in the platform." The data cannot definitively infer causality.

"In Table5, we report the multivariate regression including only the vari ables selected by elastic regression. The results indicate that 77% of variance in resilience is explained by behavioral data measured in the platform."

Discussion¶

Related Claims¶

• Behavioral grit, but not survey grit, predicts academic resilience: grittier students register smaller score declines during the pandemic — related
• Behavioral grit measured from pre-pandemic platform effort responses to setbacks predicts platform scores in math and science, especially after Covid commenced — related
• The survey grit measure has limited predictive power over platform performance in this sample and does not predict pandemic-era performance declines — related
• Student-rated teacher behaviours predict 4.39% of the variance in students' change in engagement in physical education across the first year of secondary school — related
• A machine learning model using 105 predictors predicts next-year standardized math achievement with out-of-sample prediction R-squared of .70–.78 — related
• The COVID-19 switch to remote schooling was accompanied by significant declines in math and science scores on a UAE digital learning platform — related
• Regression optimizes explanation of data at hand while machine learning optimizes out-of-sample prediction, and studies must state which aim they pursue — related