I'm comparing three groups: people that have had two infections, people that have had one infection and people that have had zero infections.
I'm showing them on a boxplot (regarding age for example) and I'm doing 3 independent two sample t-tests (group 0 vs group 1, group 1 vs group 2, and group 0 vs group 2). Am I supposed to use adjusted p values? I'm having a doubt because the variable isn't really categorical, there is an "order" 0 < 1 < 2.
Maybe it is that my idea of doing a boxplot is just wrong and I should do a linear regression between age and number of infections and use this p value instead? Should I do something else than a linear regression (in which case, what should I use??? I can't do a logistic regression because my categories "have an order")?
Is there a book / scientific article that I can cite that mentions such a situation (boxplots with categories that are ordered). I know this sounds like a standard problem but I'm struggling to use the right keywords to find stuff about it. Whenever I use the keyword order, I just have programming solutions to put the x axis in a certain oder.
