# Unsure what test to use for my non-parametric data

I would like some help choosing the most appropriate stats test to use for my data. I have two different types of organism, four different treatments and then the dependent variable is yes/no did they carry out a behavior.

For example:

EDIT


In a comment OP added: I have the raw data total number yes and total number no for each organism and treatment, which is what I used to calculate the percentage of positive responses.

• The example data contains some "positive response" which is a percentage, so it does not exactly match your description (that there would be a yes/no dependent variable - can you clarify this detail? What are you trying to find out (what scientific question are you interested in)? Commented Apr 3, 2016 at 10:19
• It looks like each row represents multiple units. Do you have the original data? Also, note that (per your title) data cannot be parametric or non-parametric; models are parametric or not. It is not yet clear whether you want a parametric model or not. Commented Apr 3, 2016 at 11:56
• Hi, I have the raw data total number yes and total number no for each organism and treatment, which is what I used to calculate the percentage of positive responses. Commented Apr 3, 2016 at 12:59
• I need to see if there are any significant differences between organisms A and B for the same treatments. Commented Apr 3, 2016 at 13:00
• Welcome to Cross Validated! Data are neither parametric nor nonparametric; those are adjectives that apply to models or techniques. If you mean "not normally distributed" that's not at all the same thing as "nonparametric" and similarly "parametric" is not at all the same thing as "normally distributed" -- one can fit parametric non-normal models and correspondingly, one can happily use nonparametric procedures on data drawn from normal distributions. Please edit your post to more clearly express the actual situation. Commented Apr 3, 2016 at 13:56

With the extra information you added in comments, It is clear you have data in binomial form, so I would start with logistic regression. Doing this in R a simple command to start could look like:
df <- data.frame(as.factor(Organism), as.factor(Treatment), Pos, Neg)