I am planning to compare Most Probable Numbers of coliforms of two samples of drinking water tested via multiple tube fermentation technique.
Isabel
. I don't know much about bacteriology or MPN techniques but it looks like there are fairly standard ways of comparing dilution series. This webpage looks like it MIGHT be a good resource. If it is and solves your problem could you let me know?
http://www.splammo.net/JLbactsite.html
Most probable number data can be a wee bit problematic. The most probable number is really the midpoint of a 'most probable range'. You can handle data like this using interval regression, where data may be represented by an upper boundary, a lower boundary or both. This allows you to use log scaling even when you have zero values in the data, since zero values are values that are in the range bounded on the upper side by the lower boundary of the first positive MPN.
Stata does interval regression very well, and it's what I have used. Cannot vouch for R, but I'd be surprised if there were not an equlivalent to Stata's intreg command.
Dear Isabel
You have two samples and each sample has many replicates ( many tubes ), now you must test your data for normal distribution .
In the case of normal distribution data you can use T-test for independent two samples, and in the case of non-normality, data you can use Mann-Whitney test to compare the two samples.
Good Luck
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis