Not sure what the appropriate test would be...

Hi,

I hope I'll be able to make myself clear.

Let's say I have 8 timepoints. At each timepoint, I randomly select a sample
of a population and calculate the proportion of people who fit into a
category.

Then an intervention happens and I do the same at another regular 8
timepoints afterwards.

By graphing all these proportions with their respective confidence
intervals, it is clear that there is a difference between those 2 groups of
8, but I would also like to be able to report a p-value.

Is it appropriate to use these proportions as continuous variables and see
if there is a difference between the mean of the first 8 and the mean of the
other 8? Or should I keep the variable as categorical and simply group all
the 8 timepoints together to perform a simple Chi-Square?

Any help or suggestion would be greatly appreciated.

Cheers,
Christophe