Hi Epid 207 Class,
The recording for this afternoon's HW7 review and the Wrap Up are now posted. Please note, the latest updated Wrap Up slides include the markings made during the "live" zoom session, and 3 new slides at the end, to clarify the last question from the review session (latest posting is 11:14 pm 3/12).
For those who did not join us live, I suggest stopping at time-point 1 hr and 33 min (approx) in the video, after we finish discussing consistency in the table, and to avoid/minimize confusion, follow slides 31-33 for a more concise answer about the last question on how to "see" exchangeability in the table (We had some technical difficulties, and things slowed down a bit).
For those of you who joined "live", I realized after class what may have been causing confusion on the last question about how to "see" exchangeability in the table. To clarify, for exchangeability to hold, you can't check for this looking just within a row for an individual, nor by looking for number patterns to match within columns or sub-columns. You need to check that certain proportions or probabilities hold, across different sub-groups (see new slide 33). What may have been causing confusion is that that particular table had rows ordered in a deliberate way, such that identical 1 and 0 sub-patterns (within the columns) could be seen, that reflected the equivalent probabilities for the exchangeability formula to be true. However, these patterns were created by a particular ordering of the rows... and some rows could have been swapped such that the 1, 0 number patterns would not match, but the math would still support the equivalent probabilities indicating exchangeability. Please see Slide 33 for further details and let me know if questions.
I also added 2 additional slides about this table, one general note, and one for consistency.
Overall, please see revised slides 31-33, in the updated deck (11:14pm time stamp).
Good Luck on your Exam!
I enjoyed having you all in the class and wish you well!
JMC