Hello Epi207 folks – Some overall comments on last week’s homework from my grading:
- Several of you are still simply forgetting to answer parts of questions. Make sure you read the whole question and answer ALL parts of a question. If you simply forget to answer parts of a question, we can’t give you credit for that part. So, please pay more attention to the question stem itself. No need to miss easy points.
- When we (and other classes) ask you to report hypothetical results or actual results from a publication, be complete. For example: What was the measure of association? What was the 95% CI? What was the p-value? Don’t continue your old habits of making generic/non-specific statements like “increased risk” or “correlated”. You’re all getting this advanced training (ATCR, Master’s or PhD) in epidemiology & biostatistics, thus you can do better & be more precise than your colleagues who aren’t getting this type of advanced training. When you go back to your home departments & are having public discussions with your colleagues, it’s that kind of precision that will separate you from your colleagues.
- When you match for variables, remember from your biostatistics training that you have to take matching into account in your data analysis plan. Several of you stated you would match for various variables, but then you forgot to mention that you would take matching into account in the data analysis phase (e.g., conditional logistic regression). See pages 171-173, “matched case-control studies”, in the “Regression Methods in Biostatistics” book by Vittinghoff et al. 2012 from your biostat 208 class.
- When you say you’re going to measure a variable, be more specific on how you would actually measure that variable. Several of you mentioned various ideas in question #5, but how would you actually measure complex variables? (for example, socioeconomic status)
Lydia & Alyssa will have more comments in your homeworks, but these are some common issues that I noticed in several homeworks.
Raj