Hello,
I have data set like attachment
where, StudySite= 3 different study location
SampleType= different fecal and environmental sample
TBX= Total E.coli count
ESBL= beta lactamage bacteria count
logTBX= log10(1+TBX) same as logESBL
ESBLProp= logESBL/logTBX
ESBLProp1= log10 (ESBL/TBX)
Hypothesis;
Effect of different study sites and parameters on bacteria count.
Analysis:
I have produce boxplot using ESBLProp and attached in the attachment. Also produce same boxplot for ESBLProp1 but log values are in minus value so hard to interpret for all audience.
Question:
1. Which process is correct to use in scientific journal article ESBLProp or ESBLProp1?
2. Please provide answer from your experience or reference?
3. Provide reference for ESBL proportion related article, I have searched it but did not get any article in which above format ESBL proportion was used.
You are not clear on what you want to do. Count data requires special methods. See the link below.
https://www.google.com/search?client=firefox-b-1&ei=QbZVW9bCHpjCjgT31baQDQ&q=regression+on+counts+R&oq=regression+on+counts+R&gs_l=psy-ab.12..33i22i29i30k1l4.157574.159121.0.162235.2.2.0.0.0.0.123.215.1j1.2.0....0...1c.1.64.psy-ab..0.1.123....0.bGIsuHavNUI
This link gives an intro. to methods and what is available in R. I think you want to know which site has more counts. Is this correct?
Best wishes, David Booth
Thanks David. Would you provide any reference for log proportion method used in any study?
- Badges
- Science topic
- Similar topics
- Journal Articles