I wonder how you would study that question (about how much freedom results in optimal happiness across the population). I'd be surprised if it evenly resolved into even thirds.