Open jonjamesjr opened 4 years ago
no one guessed the Wald story, so can keep using it replace mentimeter with clickable image if possible box of models, leader and exclusive club of best models was helpful terminology (at least to me) real life examples needed desperately better to give them coding examples where they actually do something they might actually want to do in real life? they hadn't done significance testing yet, so understanding P(we reject P) = 0.05 was hard for them. profile/conditional/marginal needed? t distribution needed? Bayes example was fine but needs a more complex one, too, real world even better confidence intervals vastly improved by thinking in terms of inverting tests, but need to make sure the link between distn of the MLE and coverage / test inversion is clear. students thought it was too long, needed more, shorter sessions they enjoyed the mentimeter interaction (so did I) more real life examples requested -- think about how to do this without getting bogged down in details some questions too big / philosophical / hard many didn't have A-Level maths found interpretation of sign. testing hard -- reorder material to include this first? breaks every hour worked well, they turned their videos back on when they were back
This year (2021) many of the students knew the Wald story.
I think many of the comments from last year still apply. Could I work in some of the finr material for these purposes, together with some actual model fitting?
Estimation.pdf Here are the student comments on Mentimeter.