Hi Class,
- Office hours on Tuesday, Feb 4 are via ZOOM conference only, 8:30-9:30.
- HW4 on PT and Bias is due Feb 13 (for a short while today, this incorrectly said 2/3 on the CLE. This typo has been fixed now.)
- HW 3 key – q 7b. After the tree diagrams, the 2x2 table for the pseudo-population was missing, though it is the same as the one obtained using the 2x2 table method beneath it. The interpretation statements have also now been made the same, across the question, as mentioned in class and to avoid confusion. We have updated this now and re-posted, in case you wish to download again.
- Clarification on Prevalent vs incident users, Slide 21 from lecture – A question was asked after class regarding why Subject A is not a prevalent user. To clarify, we started off by saying we got rid of all prevalent users (first bullet)… and then the colored lines indicate what actually happened after Surgery (point of eligibility), some people started using aspirin (like person A) and some didn’t (person B). I brushed over in class about how we got rid of prevalent users and just said "let’s assume this". Perhaps in practice, you would interview participants at the point of surgery, when assessing eligibility and ask them if they were using aspirin regularly (or give them a “baseline survey).. those who were, would not be in your analysis population for this study. Then you send a follow-up survey, and it takes time to get this back. I think the rest of the scenario can proceed as described in the note.
- Dr. Graff will be lecturing live on DAG’s on Feb 6, next week.
Have a good week!
JMC