ruizt / stat218 Goto Github PK

check lifetime units (days vs weeks??); include confidence level on tables

inconsistency: hypotheses on slides and lab for cholesterol example differ

apparently, package conflicts wrt q5

Typo: interval slide gets directions flipped in ci formulae
Add: pairing can also come from multiple units

notation is sloppy, needs fixing: (a) subscripting and argument conventions differ between OR and RR; (b) population quantities and estimates get mixed in some of the formulae for SEs, etc.

remove x axis on t plots, line confuses students
switch shading on t plots

Table orientations still confusing, consider using standardized names (eg case/control/exposed/unexposed) and differentiating study designs clearly

consider changing 'paired two-sample' and 'independent two-sample' to 'paired differences' and 'two-sample'; having two types of two-sample problems seems to confuse students

error: 'accuracy' in place of 'precision'

OR/RR distinction should be made in terms of sampling constraints, not prospective/retrospective or other design terminology

Error in slide on percentiles — need to specify whether value is included with above or below values in “vanilla” example… not consistent with median definition as is

hiring dataset is currently a data frame, needs to be coerced to a table

second portion of lab not sequenced well in slides

have extra lecture time; either add examples or include measures of association

se formula should have individual se's squared

include example of how to separate by group

takes too long to re-render site and pdf copies are not in use in class

on comparisons with control slide, "increases lifespan" should read "increases mean lifespan"

consider adding more example intervals and statements at end

add midpoint explanation to slides

missing material after test 2