I have a smoothbrain statistics problem for /sci/.
Let’s compare a vaxxed and unvaxxed population. Of a defined population of people (1500), 80% (1200) are vaxxed and 20% (300) unvaxxed. Unvaxxed are tested weekly, vaxxed only when symptomatic. The vaccine is non-sterilizing. By the end of three months there are 50 positive cases in the vaxxed group and 35 in the unvaxxed group. Does this mean the unvaxxed group is almost 3X more likely to catch the virus?
The vaxxed group is undersampled and each unvaxxed positive is biased as it counts much more to the postive/negative ratio.
The underpinnings of the statistical test are faulty but mathematically why is this wrong? What is the proper test and statistics to report here?
Let’s compare a vaxxed and unvaxxed population. Of a defined population of people (1500), 80% (1200) are vaxxed and 20% (300) unvaxxed. Unvaxxed are tested weekly, vaxxed only when symptomatic. The vaccine is non-sterilizing. By the end of three months there are 50 positive cases in the vaxxed group and 35 in the unvaxxed group. Does this mean the unvaxxed group is almost 3X more likely to catch the virus?
The vaxxed group is undersampled and each unvaxxed positive is biased as it counts much more to the postive/negative ratio.
The underpinnings of the statistical test are faulty but mathematically why is this wrong? What is the proper test and statistics to report here?
