This report presents a comprehensive Multivariate Analysis of Covariance (MANCOVA) examining student academic performance across math, reading, and writing scores while accounting for demographic and academic factors such as gender, race or ethnicity, lunch type, and test preparation. Two modeling approaches are compared, including a full factorial MANCOVA to capture complex interactions among variables and a simplified one way MANCOVA to isolate the effect of gender while controlling for socioeconomic background through parental education. The analysis includes assumption testing, effect size estimation, post hoc comparisons, and visualization, providing insights into the relative importance of demographic and educational factors in shaping academic outcomes as well as highlighting potential areas for intervention such as test preparation programs.

A working file for the training in AMMnet

Updated version with different synthetic corpus data, tuned features. Still needs analysis better capturing word context and relevant dictionaries.

assignment