It really depends on what the research is on, but from my experience...
If "experimental" means animal models: better overall picture (at least for me since most of our stuff is done in vitro first) , less tedious (it's more interactive), less chance of developing tendinitis (less pipetting), more frustrating (animals behave less consistently than cells), more time consuming (setup/clean up/maintenance), less control over schedule (harder to control when they are of a certain age, pregnant, etc, ie I've spent Thanksgiving and Christmas in the lab due to mice), need more planning (can't grow a mice colony overnight), can be harder to learn procedures (but once learned can be applied across species), more involvment even at a basic level (even with ugrads)
"Molecular" projects: more tedious (especially if you do cultures), tendinitis (pipetting a 384well plate is serious business), easy place to put ugrad "muscle" (labs need someone to pour gels or run routine PCRs), quicker turn-around per experiment, easier data collection (in general), easier data analysis (in general), easier to fit around schedule, usually cheaper (good for both number of experiments and covering for screwups), more respected (debatable, but sometimes animal models get the shaft, they are usually the last experiment done before submitting a paper), techniques are fairly standard (makes switching labs easier)
obviously both have their good points and bad points, a lot depends on the project and your personality. Also consider whether you think you want to continue research in the future, if you don't I think you get "more" out of animal models in a shorter amount of time.