search

charlesfrye / AppliedStatisticsForNeuroscience

Materials for UC Berkeley Neuroscience 299
52 stars 17 forks source link

Lab 06 Issues #11

Closed charlesfrye closed 6 years ago

charlesfrye commented 7 years ago

~Definition of omnibus test needs to also be in part b~

Comment omnibus test code out from student version and indicate that it's for later.

charlesfrye commented 7 years ago

~Definition of omnibus test needs to be clearer in part a~

Moving omnibus test discussion into part b, and making part b the only part.

charlesfrye commented 7 years ago

Definition in part b is incomplete -- describes how to get the effect term, but not the error term.

EDIT 08/09/2017 - can't determine what this was about.

charlesfrye commented 7 years ago

There should be a way to simulate the "one strong effect" case for the final question on omnibus testing. That way students who don't get the question correct can still see the results.

charlesfrye commented 7 years ago

Reorganizing this:

The tutorial will contain the material from Lab A, on implementing 2-Way ANOVA by hand and on interpreting interaction plots (remember the comment from Fisher that ANOVA is to be used to determine whether patterns we see in graphs are due to chance or not). ~Will need to generate more synthetic data from alternate outcomes of the anxiety reduction experiment~. Chose to use totally fake data instead.

The lab will focus on multiple comparisons and ANOVA, explaining family-wise error rate and omnibus tests. Would be useful to also spend some time on the "avoiding multiple t-tests" motivation for using ANOVA in the previous lab.

charlesfrye commented 6 years ago

Added post-hoc t-tests to previous lab. Kept material on interaction plots. Made Lab A primarily about visualizing data and using statsmodels to perform ANOVAs.