Hi,
I am looking for some guidance on the appropriate statistical analysis for part of my study. I am looking at outcomes related to infection among 7 different groups based on source of the infection. I want to compare significance between all groups. For example:
Source 1: 2 Dead, 15 alive
Source 2: 9 Dead, 12 alive
Source 3: 9 dead, 5 alive and so on for 7 different groups.
What would be my best test to run? Thank you.
Does your theory have any predicted ordering of the mortality of these, or better yet a quantitative prediction? E.g., is source the dose of something, or amount of exposure? How was the sampling done? Are the data iid (independent and identically--maybe Bernoulli--distributed)? Are you interested in just whether there are differences because of the source? It sounds like you want all pairwise comparisons (7 choose 2 is 21, and given the sample sizes the power may be too low). Depending on your answers to these, logistic regression may be useful, but it will depend on your answers.
Daniel Wright no predicted ordering of mortality. The source is referring to the nidus of infection. These are all patients with a specific bloodstream infection from a single hospital over the course of about 5 years, but the groups are the breakdown based on source of said bloodstream infections (e.g. pulmonary source, urinary source, wound infection, etc.)
The Appropriate test for comparison between each two groups is Chi square test. For the comparison among all groups kruksal walas test can be used to give a result that there is a significant difference among the groups.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis