# What is the appropriate statistical test for a choice chamber experiment?

The Design of my experiment involved a choice chamber (see below) with two sides ( one with a black background and one with a white). Then an ant was placed in the middle and timed how long it spent in each side. We repeated this many times. This was to establish if the ants showed a preference of background. We then placed cotton balls on each side with different sucrose solutions (0%, 25% or 75%). We then did different combinations of these conditions so see if the ants would change their preference for a greater reward (i.e. a higher sucrose solution).

I am trying to determine the best way to analyse this data. I did have it with the time spent on each side but then decided I didn't need this as the time spent in condition 2 is the same as total time - time spent in condition 1. So now I just have the time spent in condition 1 and the factors for that condition (The background and sucrose concentration). The data is not normally distributed so looking at non-parametric tests. I think I should be using a Mann-Whitney U test. (I am using this to help make my decision and lots of thought https://dzchilds.github.io/stats-for-bio/images/stats_key.pdf )

Any help/thoughts would be really appreciated? I am really looking for general answers if anyone knows a good statistical test for use with choice chambers.

Choice Chamber Example (https://i.stack.imgur.com/dMsTf.png)

Thanks,

Robert Wilson