Binomial and Nominal data analysis question

#1
Hi everyone,

This is my first post here! I am a novice at stats and could use some guidance. I am working in JMP pro. I have a data set where 1 column is tree type (lets say type 1, 2, or 3) and the other column is binomial (1,0) dead or alive. I am trying to get the statistical significance for the difference in average % mortality by tree type.
Basically I need the hard stats to explain the figure below:
 

Karabiner

TS Contributor
#2
I am trying to get the statistical significance for the difference in average % mortality by tree type.
Well, you don't compare averages here, but proportions.
Which test would be appropriate, depends on your sample size and what exactly you want to know (difference in
general? pairwise comparisons?). The usual approach would be a 3x2 table with Chi² test.
Basically I need the hard stats to explain the figure below:
I am not sure what you mean by "hard stats", or by "explain".

With knd regards

Karabiner
 
#3
Thank you for the reply! For clarification I need to know whether or not the difference in % mortality between the tree types is significant. (Ideally how to do it in JMP, but anything helps!) I have 38 rows of data with with one column being 1 or 0 for "alive" or "dead" and the other columns being tree type "1, 2, or 3".

Thank you for any help!
 

Karabiner

TS Contributor
#4
So a Chi² will probably run into problems, since you only have 38 observations,
distributed across 6 cells, and you have extreme outcome distributions
(about 90:10). This will violate the assumption that not more than 1 cell here
should have an expected (!) frequency < 5.

If JMP offers an exact test (Fisher's exact test) for tables larger than 2x2,
you should use that instead of Chi².

If JMP offers Fisher's test only for 2x2 tables, you can do 3 pairwise
comparisons.

If JMP does not offer Fisher's test, then you can perhaps look for an online
calculator.

With kind regards

Karabiner