Which statistics test to use?

#1
Dear all,

I am working on a project about a diagnosis system where I have two scenarios.

In the first I have binary data like:
highTemp
0
1
etc.
where zero indicate normal temperature and 1 indicate high temperature. I want to correlate this with a class whose values can be one of three mechanic systems (dependent variable) that I encoded 1, 2, and 3. Which test should I use in this case? I thought about chi square but my data (highTemp) is not normally distributed.

In the second problem, I just wanted to know how likely is to see an independent binary variable like highTemp occur when we have another independent binary variable like engine speed (encoded as 0 for normal speed and 1 for too fast). Here as well, my both binary variables are not normally distributed according to Shapiro-Wilk test. Would Chi-square work here? Chi-square is the only test I could see that does not require a dependent and independent variables, but it does require normal distribution though.

Your help is highly appreciated
 

Karabiner

TS Contributor
#2
You have categorical data, therefore considerations of normality do not apply.

Chi² seems reasonable here.

With kind regards

Karabiner
 
Last edited:
#3
Dear Sir/Ma
Group C is a control group. No surgeries to the group but sampled as the other 3 groups that had different techniques of a rumenotomy procedure. Each group had 3 males and 3 females as sub group. Sampling were before surgery as PRE,then postsurgeries as 0hr, 5hrs, 24hrs,48hrs, 72hrs, weeks 1,2 and 3. How do I compare the groups. I'm using graphpad prism 5.
Assays were for Cytokines.

Thank you.