There are many arguments presented for and against learning abstract concepts as opposed to real world applications. A study done by Ohio State scientists indicates that abstraction might be better, at least under certain contexts.
The trick is to learn how to figure out the rules of the game. The notion behind abstraction is to learn the underlying principles rather than how such principles manifest in the real world. In contrast, learning via examples, not knowing the mechanics of abstraction, provides the nice "aaha" moments, without an understanding of the underlying principle. It is a valid argument that in certain sciences you need to look, touch, feel and play with things to understand them better, whereas others are more amenable to learning via abstraction.
A reflection on my own learning during my undergraduate education in engineering tells me that I was better at abstracted learning rather than learning by playing with things in the laboratories. I scored very nicely in theoretical courses with equations and derivations while losing out in the laboratories :). I am not sure if I can conclude that I was better learning the principles from my scores, but for some reason I was better at manipulating equations rather than play with the diesel engine in the laboratory.
People seem to have an intrinsic preference for one type of learning as opposed to the other. Or maybe they use a combination of the two to achieve optimality. What does become important however, is while teaching kids at home, we might have to pay attention to the principles as well, and not just be happy seeing them play with toys that provide the "touchy" "feely" learning.