I have a pretty general question: I'm writing on my thesis and have some linear regressions that include a continuous variable. Is it okay to use the variables as continuous for the linear regression and use the same one as categorical (2 groups via median split) to analyze the means with post hoc(for example via GLM on SPSS)? I'm aware that it's not the best solution, but I don't know how to look at means otherwise.
It is generally bad practice to categorize continuous variables; it reduces power, increases type I error and introduces "magical thinking" - i.e that something amazing happens at the median.
Why do you even want to "look at the means" of the groups based on median split? Why median? Why two groups? Why not three or ten or whatever?
Unless you have some strong theoretical reason why there should be a difference at the median, I think your aim is misdirected. Regression lets you see if the continuous variable is significant; you can add quadratics or splines if you want to look at nonlinear effects. And you can compute the expected mean at any level.